An efficient overset grid technique for computational fluid dynamics based on method coupling and feature tracking

A new overset grid method that permits different fluid models to be coupled in a single simulation is presented. High fidelity methods applied in regions of complex fluid flow can be coupled with simpler methods to save computer simulation time without sacrificing accuracy. A mechanism for automatically moving grid zones to track unsteady flow features complements the method. The coupling method is quite general and will support a variety of governing equations and discretization methods. Furthermore, there are no restrictions on the geometrical layout of the coupling.;Four sets of governing equations have been implemented to date: the Navier-Stokes, full Euler, Cartesian Euler, and linearized Euler equations. In all cases, the MacCormack explicit predictor-corrector scheme was used to discretize the equations.;The overset coupling technique was applied to a variety of configurations in one, two, and three dimensions. Steady configurations include the flow over a bump, a NACA0012 airfoil, and an F-5 wing. Unsteady configurations include two aeroacoustic benchmark problems and a NACA64A006 airfoil with an oscillating simple flap. Solutions obtained with the overset coupling method are compared with other numerical results and, when available, with experimental data. Results from the NACA0012 airfoil and F-5 wing show a 30% reduction in simulation time without a loss of accuracy when the linearized Euler equations were coupled with the full Euler equations. A 25% reduction was recorded for the NACA0012 airfoil when the Euler equations were solved together with the Navier-Stokes equations. Feature tracking was used in the aeroacoustic benchmark and NACA64A006 problems and was found to be very effective in minimizing the dispersion error in the vicinity of shocks.;The computer program developed to implement the overset grid method coupling technique was written entirely in C++, an object-oriented programming language. The principles of object-oriented programming were exploited to simplify the program design. The success of the program demonstrates that object-oriented languages may be superior to procedure-oriented languages for some scientific computing applications.